#!/usr/bin/env python
# coding: utf-8

# # Show, Attend and Tell: Neural Image Caption Generation with Visual Attention

# * 바벨피쉬 / 딥엘라스틱 : 파트 4 - 딥paper [1]
# * 김무성

# # 참고자료
# * [1] Show, Attend and Tell: Neural Image Caption Generation with Visual Attention - https://arxiv.org/abs/1502.03044
# * [2] Recurrent Neural Network (RNN) 과 LSTM - http://www.whydsp.org/280
# * [3] (slideshare) Show, Attend and Tell: Neural Image Caption Generation with Visual Attention - http://www.slideshare.net/eunjileee/show-attend-and-tell-neural-image-caption-generation-with-visual-attention
# * [4] (slideshare) Attention mechanisms with tensorflow - http://www.slideshare.net/KeonKim/attention-mechanisms-with-tensorflow
# * [5] (code) kelvinxu/arctic-captions - http://kelvinxu.github.io/projects/capgen.html    
# * [6] (code) DeepRNN/image_captioning - https://github.com/DeepRNN/image_captioning
# 
# #### Data
# * Flickr8k - http://nlp.cs.illinois.edu/HockenmaierGroup/8k-pictures.html
# * Flickr30k - http://web.engr.illinois.edu/~bplumme2/Flickr30kEntities/
# * MS COCO - http://mscoco.org/
# * MS COCO API - https://github.com/pdollar/coco